The AI market is still there, but stock prices have already fallen.

On November 5 local time, the U.S. stock market experienced a severe sell-off. According to BiyaPay's data, the Nasdaq index dropped 5.52% in one day, closing at 24,315 points; the S&P 500 index fell 4.36%, and the Dow Jones Industrial Average decreased by 3.44%. Technology stocks and the semiconductor sector were the main forces behind the decline, with the Philadelphia Semiconductor Index dropping over 12%.

The performance of individual stocks was even more dramatic. SanDisk (SNDK) fell 22.26% in one day, Micron Technology (MU) dropped by 17.46%, and Western Digital (WDC) decreased by 17.35%; AMD dropped 17.26%, and NVIDIA also fell 9.26%.

The most concerning aspect of this decline is not just how much it fell, but the scope and direction of the market sell-off: from chip design and storage devices to data center-related companies, multiple links in the AI industry chain are under pressure simultaneously.

The market's extreme volatility has also made investors more concerned about capital arrangements between different markets. For users focusing on both digital assets and traditional financial markets, how to connect USDT with U.S. stocks, Hong Kong stocks, and other markets is a question worth considering in global asset allocation. The global one-stop asset allocation platform BiyaPay supports multi-currency exchanges and provides investment markets such as U.S. and Hong Kong stocks. Users can use USDT as capital entry, exchange it for U.S. dollars or Hong Kong dollars within the platform, and then make investment arrangements according to the corresponding accounts and trading rules. This type of capital connection helps reduce operational steps between different platforms, making multi-market capital management clearer.

Previously, the market was trading on expectations of continuous growth in AI demand, tech giants increasing capital expenditures, and the accelerated expansion of data centers. Now, investors are starting to recalculate another account: can these inputs be implemented on time? Will infrastructure construction be limited by policy and energy conditions? How much profit can massive capital expenditures eventually be converted into?

This may be the more significant change behind this sharp drop in U.S. stocks.

Why did the Nasdaq suddenly drop? The midterm elections are just a trigger; concerns over AI investments have long existed.

The sell-off in U.S. stocks on November 5 is closely related to changes in policy expectations following the U.S. midterm elections, but attributing this drop entirely to the elections underestimates the pressure that has already accumulated within the market.

Before the election results were announced, the AI sector was already facing several controversies: Tech giants were continually increasing capital expenditures; could AI commercialization revenue keep up? The construction of data centers requires significant electricity and infrastructure; is the project return cycle too long? Have the high growth expectations for chip companies already been fully reflected in their stock prices?

These issues did not just emerge on November 5; the market had previously been willing to grant a higher tolerance for valuations.

After the announcement of the midterm election results, the Democratic Party achieved significant electoral success, prompting the market to reassess the future policy environment for AI and data center construction in the U.S. Some local elections and referendums also brought the land use, energy consumption, water use, and community impact of data centers to the forefront.

The market fears that some projects may face stricter approvals, longer construction cycles, or even pauses or adjustments in the future. It is important to note that policy discussions and local constraints do not equate to the U.S. having completely halted data center construction. Different states and localities have different policy authorities and project conditions, and it cannot simply be understood as a complete halt to data center construction.

However, capital markets often trade expectations rather than waiting for policies to materialize.

For the AI industry chain, even if only some projects are delayed, it could affect the order rhythm and income recognition of related companies in the coming quarters. Previously regarded as highly certain growth expectations, any emergence of uncertainty could directly impact valuations.

More importantly, the market had already been facing issues of crowded AI trading, high valuations, and doubts about capital expenditure returns before the elections. The election results acted more like a catalyst, concentrating the worries that had originally accumulated.

Why are AI storage stocks dropping harder? The market is worried about demand, prices, and profit cycles simultaneously.

If the overall decline of the AI industry chain reflects a decrease in market risk appetite, then the sharp drop in storage stocks also involves the cyclical characteristics of the industry itself.

AI model training and inference require a significant amount of data processing capability. GPUs handle calculations, HBM provides high bandwidth memory support for processors, and DRAM and NAND undertake different levels of data storage and read/write tasks. As the scale of AI servers and data centers expands, the importance of storage devices also increases.

This has been one of the reasons storage companies have received capital attention in recent times.

However, the storage industry does not simply see profits grow with increased demand. It is also affected by factors such as supply expansion, product prices, capacity utilization, client inventory, and product structure.

This adjustment can be understood from three levels.

First, increasing AI demand does not mean that all storage products can benefit synchronously.

HBM, server DRAM, enterprise-grade SSDs, and traditional hard drives face different client demands, technical barriers, and supply and demand environments. Even if AI data centers continue to expand, the order growth and profitability of different products might diverge.

Second, storage prices and profit margins are cyclical.

When industry supply is tight and product prices rise, storage manufacturers' profit capabilities might improve rapidly; however, if companies expand production, supply increases, or client procurement rhythms change, product prices and profit margins could face pressure.

What the market is really worried about is whether the current high-profit levels can be sustained, not just how many orders remain today.

Third, the market may trade changes in future profit expectations ahead of time.

Stock prices reflect future expectations rather than the current operational status of the companies. If investors believe that product prices, profit margins, or order growth may fall below prior expectations in the coming quarters, even if the company's current performance remains strong, the stock price may decline in advance.

Although they are all storage stocks, what are the differences in the decline logic of Micron, SanDisk, and Western Digital?

The simultaneous significant drops of Micron, SanDisk, and Western Digital easily lead one to regard them as the same type of AI storage asset. However, from a business structure perspective, the three companies are not entirely the same.

Micron Technology primarily operates DRAM and NAND flash memory products, with HBM being an important component of AI servers. The demand for high-performance memory driven by the expansion of AI data centers is a critical reason for the market's attention to Micron.

SanDisk mainly focuses on NAND flash memory and related storage products, and its business is closely related to flash demand, product prices, and supply-demand cycles.

Western Digital, on the other hand, primarily focuses on hard drive business. As the data volume in data centers continues to increase, enterprise-grade hard drives still have application space, but its product structure and profit-driving factors differ significantly from Micron's DRAM and HBM business.

Therefore, one cannot simply categorize the three companies as HBM suppliers, nor assume that they will show completely identical fundamental performance due to the same changes in AI demand.

This is also a point that requires further segmented research after the sharp decline in the sector: Micron needs to look at the competitiveness of its HBM and DRAM businesses, supply, and prices; SanDisk needs to focus on NAND product prices and inventory; and Western Digital needs to pay attention to enterprise-grade hard drive demand, product structure, and profitability.

Sector sentiment can change synchronously, but the companies' fundamentals may not align.

Will data center construction really cool down? Industry data shows another side.

The market's concerns about delays in data center projects do not mean that AI infrastructure demand has disappeared.

CBRE's report on North American data centers for the first half of 2026 shows that in the eight primary North American data center markets it tracks, the in-construction capacity reached 7,481 megawatts, a 24.8% increase compared to the previous report period; of this, over 80.4% of the in-construction capacity has already secured pre-leasing commitments.

The report also indicates that the total supply in these primary markets has reached 10,903 megawatts, with a vacancy rate falling to 1.4%. This reflects that data center capacity remains tight, and the demand for infrastructure from AI and large cloud service providers remains strong.

This set of data indicates that data center construction has not lost its demand base due to market worries.

However, it also reveals another issue that is easily overlooked: strong demand does not mean all projects can be delivered as planned.

Data center construction requires not only land and funds but also stable power supply, supporting facilities, approvals, and adequate construction capacity. CBRE's report also points out that power availability, infrastructure delivery cycles, and opposition from local communities are becoming important factors influencing project site selection and construction.

This suggests that the data center industry may see a more pronounced differentiation in the future: projects that can secure power resources, complete approvals smoothly, and deliver on time still have the chance to benefit from robust demand; meanwhile, those with inadequate power conditions, approval hindrances, or excessive financing costs may face delays and rising costs.

Thus, the real focus should not be on whether there is demand for data centers, but whether that demand can be smoothly converted into deliverable capacity and operational revenue.

Micron's historical price-to-earnings ratio is only about 8.8 times; why did its stock price still suffer a sharp decline?

During this adjustment, Micron's valuation is also worth discussing separately.

According to Micron's reported FY 2026 results, the company's annual revenue reached $133.188 billion, with a GAAP diluted earnings per share of $74.33. Based on the closing price of $656.39 on November 5, the historical price-to-earnings ratio is approximately 8.8 times.

Looking solely at this number, Micron does not seem expensive. However, the problem is that the storage industry is a typical cyclical industry, and one cannot solely use current profits to determine whether the valuation is reasonable.

When the industry is in an upturn, with rising product prices and improving capacity utilization, company profits may grow rapidly. At this time, the price-to-earnings ratio calculated based on high profits often appears relatively low.

However, if the market anticipates future product price declines and narrowing profit margins, changes in corporate earnings might occur, and even if the current price-to-earnings ratio is low, investors might preemptively adjust valuations downwards.

This is a common phenomenon for cyclical stocks: at their most profitable moments, price-to-earnings ratios may appear the lowest; yet, what the market is truly worried about is whether the current high profits can be sustained.

Therefore, Micron's price-to-earnings ratio of about 8.8 times is based on already reported historical earnings for FY 2026, not future earnings forecasts, and cannot directly conclude that its stock price is cheap.

What is more noteworthy is whether the company's HBM shipments, DRAM and NAND prices, gross margins, capital expenditures, and customers' long-term procurement arrangements can continue to support its profit levels in the coming quarters.

Not just changes in fundamentals; how does concentrated withdrawal of funds amplify the decline?

Another factor that cannot be overlooked is that this drop may not only be a result of investors reassessing corporate value but also involves concentrated withdrawals of funds from popular trades.

Previously, the AI industry chain had experienced a prolonged rise, wherein some tech stocks had high valuations, and market trading was relatively crowded. When policy expectations, performance concerns, and interest rate environments change simultaneously, investors might choose to reduce their risk exposure.

If institutional investors adjust their positions simultaneously, or if certain trading strategies require reducing leverage, selling pressure could appear concentrated over a short time.

This type of capital behavior can amplify price volatility, especially in stocks that have had significant gains and where market positions are concentrated.

However, it is important to distinguish that a rapid decline in stock prices does not directly prove that there has been large-scale forced liquidation in the market, nor can it be inferred that certain funds are experiencing liquidity crises. Determining whether capital has undergone passive deleveraging still requires considering transaction volumes, financing data, fund holdings, and derivative markets.

For ordinary investors, it is crucial to recognize that short-term stock price fluctuations may not be entirely determined by changes in corporate fundamentals; market sentiment and capital structure may also accelerate the speed of decline.

What to watch after the crash? Whether AI storage stocks can stabilize depends on these key indicators.

After this significant adjustment, rather than rush to judge whether the AI market has ended, it may be more valuable to focus on several key indicators.

First, observe the capital expenditure plans of tech giants. If large cloud service providers continue to maintain investment in AI infrastructure, it indicates that long-term demand still has support; if companies begin to adjust budgets, delay projects, or change construction rhythms, then the growth expectations for the industry chain need to be reassessed.

Second, observe the prices and gross margins of storage products. For companies like Micron and SanDisk, product prices, order structures, inventory, and capacity utilization often reflect industry prosperity better than daily stock performance.

Third, pay attention to the actual delivery status of data center projects. Pre-leasing and planning can only indicate potential demand; whether they can ultimately secure power, complete construction, and enter operations determines whether related investments can generate actual income.

Fourth, focus on AI commercialization and cash flow. Whether AI companies can achieve sustained income through model services, cloud computing, software, and enterprise applications will directly affect the market's acceptance of high capital expenditures.

Fifth, observe whether capital flows back into the tech sector. If subsequent market rebounds are driven only by a few heavyweight stocks, while the storage, semiconductor, and data center industry chains remain weak, it indicates that the divergence in capital regarding AI investment logic may not have ended.

Is the AI bull market over? The market may be entering a phase that values profits more.

The stock market crash on November 5 cannot be simply understood as the peak of AI industry development; nor can it be concluded that this adjustment is merely a short-term sentiment issue just because data center demand remains strong.

A more reasonable interpretation is that the market is reassessing the quality of growth in AI investments.

In the past, investors focused more on the growth of computational power demand, data center expansions, and capital expenditures by tech giants; now, the market is beginning to question whether projects can be implemented on time, whether companies can continuously secure orders, whether product prices and profit margins can be maintained, and whether substantial investments can ultimately be converted into cash flow.

This means that the AI industry chain may no longer experience a market where all companies rise together, but rather progressively enter a stage that emphasizes realized profits, product competitiveness, and capital returns.

For storage companies like Micron, SanDisk, and Western Digital, long-term demand remains worth tracking, but whether short-term stock prices can stabilize will ultimately depend on whether orders, prices, inventory, and profits can withstand market scrutiny.

The long-term story of AI may not have ended, but the phase where telling growth stories alone can drive stock prices up may be undergoing changes. Moving forward, what the market truly needs to verify is who can turn AI demand into sustained, realizable profits.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。